Identification of PfPP5 cDNA and gene

To identify new Ser/Thr phosphatases of Pf, we have recently undertaken a PCR-based approach. At first, we made degenerate deoxyoligonuclotide primers corresponding to the conserved peptide sequences GDXHGQ and GDXVDRG of PPs [17]. An approximately 120 bp PCR product, obtained by using these primers with the Pf 3D7 genomic DNA as template, was cloned in the pGEM-T vector (Promega) using the T/A cloning strategy. Sequencing of individual clones revealed a variety of potential PPP sequences in Pf, one of which was an exact match with a putative open reading frame (ORF) on Pf chromosome 13 in the sequence database of the Sanger Center (Accession No. AL049185). Comparison of the predicted primary structure of the protein clearly revealed its similarity with known members of the PP5 family, including the presence of multiple TPR domains (Fig. 1 and 2). Specific primers were then made against this ORF and the 1578 base pair gene was cloned and confirmed by sequencing. The same set of primers were then used to amplify the PfPP5 cDNA sequence from a poly(A)-enriched mRNA population of Pf strain Dd2. This sequence fully matched the genomic Sanger sequence, suggesting the lack of introns in the PfPP5 gene.

Recombinant expression of PfPP5

To express the putative PfPP5 protein, we subcloned it in the bacterial expression vector pET-15b and introduced the plasmids into E. coli BL21(DE3). As shown in Fig. 3, when induced with IPTG, the pET-15b-PfPP5 clone produced a polypeptide of approximate Mr 70 k (lane 2), which is in excellent agreement with the calculated MW of 69,139, considering that pET-15b adds a (His)₆ tag (about 2 kDa) to the N-terminus of the recombinant protein. The presence of the (His)₆ tag could be confirmed by anti-His antibody in immunoblot (data not shown). The His-tagged PfPP5 was subsequently purified through standard nickel-chelation chromatography (Fig. 3, lane 5). Using ³²P-labeled histone as substrate under standard assay conditions (in absence of any activator), the specific activity of the recombinant PfPP5 was measured to be 198 ± 25 nmoles of ³²P/min/mg enzyme.

The deletion mutant, ΔTPR-PfPP5, was similarly cloned by PCR using a primer corresponding to the internal sequence, and purified by nickel chelation affinity chromatography. The deletion mutant starts with Met-273, as described in Fig. 1.

Identification of native parasitic PP5

To identify whether PfPP5 is actually expressed in the parasite, cell-free extracts of the asynchronous erythrocytic stage parasites were subjected to standard chromatographic procedures that had been previously optimized for human PP5 [18]. Studies of mammalian PP5 have revealed that the TPR domain might act as an auto-inhibitory domain of the phosphatase activity, and that tryptic digestion of PP5 causes loss of the TPRs and concomitant stimulation of activity [19–21]. As described under Materials and Methods, the purification was in fact monitored by assaying for a trypsin-activated phosphatase with the assumption that PfPP5 may behave like mammalian PP5 in this regard [18, 20, 21]. As mentioned, the activity peaks always coincided with a ~69 kDa band in parallel Western analyses of the chromatographic fractions. In SDS-PAGE, the partially purified fraction contained a major polypeptide of the expected size (69 kDa) (Fig. 4, lane 2). In Western blot, the rabbit antibody raised against the purified recombinant protein also specifically reacted with the native polypeptide (Fig. 4, lanes 5, 6). A similar fraction prepared from uninfected erythrocytes contained very little protein (Fig 4, lane 3), and did not react with the antibody (lane 7). A pre-immune serum served as a negative control (Fig 4, lane 9). As expected, the native Pf protein was about 2 kDa smaller than the His-tagged recombinant (compare the native band in lanes 2, 5, and 6 with the recombinant in lanes 4 and 8). Using phosphohistone as substrate, the native Pf enzyme exhibited a specific activity of 210 ± 20 nmoles of ³²P per minute per mg enzyme, which is comparable to that of the recombinant enzyme. Together, these results demonstrated the equivalence of the native and recombinant proteins, and in turn, confirmed the authenticity of the cDNA sequence.

Primary structure of PfPP5

When the predicted primary structure of PfPP5 was aligned with known PP5 sequences (Fig. 1), the following features were obvious. First, PfPP5 contained the catalytic core found in all Ser/Thr phosphatases of the PPP family, including the signature motifs such as GDXHGQ, GDFVDRG, RGNHE, HGLL, and SAPNYCD, to name a few [22, 23]. Site-directed mutagenesis and structural studies in PP 1 and PPλ have previously established the roles of specific amino acid residues in these domains in the various aspects of catalysis, such as metal ion binding, phosphate recognition, and co-ordination of water molecules [24–26].

While the catalytic core is generally necessary and sufficient for the phosphohydrolase activity of PPP enzymes, the residues outside the core play critical roles in binding accessory proteins or small molecules that modulate the catalytic activity [25–29]. Specifically in the PP5 class, the N-terminus has been shown to contain three TPR motifs [5–8], the three-dimensional structure of which is now also available [8]. Interestingly, the N-terminal sequence of PfPP5 was the longest of all, and seemed to contain four TPR motifs instead of three (Fig. 2). Thus, in comparing the full sequences of the PP5s, only three of the four PfPP5 TPRs were aligned with the others, and the second TPR of PfPP5 was left out. A detailed sequence analysis of all the PfPP5 TPRs is offered below.

The TPR motif is a degenerate, 34-amino acid repeat that is often found in tandem arrays, sometimes separated by spacer sequences [9]. We propose such a 22-residue spacer between the first two TPRs of PfPP5 (Fig. 1). Although no single amino acid is absolutely invariant in all TPRs, they do contain a largely conserved pattern of amino acid similarity in terms of size, hydrophobicity, and spacing [8, 9]. Eight amino acid residues are critically placed on the same face of their respective helices: 4, 8, and 11 on the first helix, and 20, 24, and 27 on the second [8, 9]. As shown in Fig. 2, residues that are identical or have similar hydrophobicity are found in all four proposed TPRs of PfPP5 in the correct relative spacing. Moreover, secondary structure prediction suggested that each PfPP5 TPR also has the potential to form the two conserved helices (A and B), as marked in Fig. 2. Thus, all the TPRs of PfPP5 may satisfy the structural requirements of a TPR motif.

An approximately 34-residue stretch following the last TPR of PP5 (Fig. 1) shows weak similarity to a TPR motif. However, alpha-helix prediction and structure determination have shown that it is a single long helix that extends out of the TPR domain [8]. Thus, this region may not represent a typical TPR and therefore, is simply marked as "helix" in Fig. 1.

Auto-inhibitory role of PfPP5 TPR: activation by polyunsaturated fatty acids

The biochemical properties further confirmed the PP5-like nature of the Pf enzyme. First of all, members of the PP5 class are distinguished by their profound stimulation by polyunsaturated fatty acids [19–21, 30]. As shown in Fig. 5, the phosphatase activity of PfPP5 increased with increasing concentrations of arachidonic acid and oleic acid. At its highest, the fatty-acid stimulated activity was about 3 times the basal activity and was exhibited at approximately 70 and 80 μM of arachinodic and oleic acids, respectively. In contrast, stearic acid, a saturated fatty acid, did not appreciably stimulate PfPP5 activity.

To test the role of TPR region in this activation, a TPR deletion mutant of PP5 (ΔTPR-PfPP5) starting at the Met-273 was expressed with an N-terminal (His)₆ tag. The recombinant protein was expressed from pET-15b clone and purified by Ni-chelation chromatography (Fig. 3, lane 6). In SDS-PAGE analysis, ΔTPR-PfPP5 exhibited a Mr of approximately 40 k as expected (predicted size of 37,098 for the PP5 part plus about 2,000 for the His₆ tag). Interestingly, the specific activity of the ΔTPR-PfPP5 enzyme was about 3-times that of the full-length PfPP5 without arachidonic acid (580 ± 35 nmoles of ³²P per minute per mg enzyme; Fig. 5). Moreover, unlike the full-length enzyme, ΔTPR-PfPP5 was not activated by arachidonic acid, suggesting that the TPR region is required for the activation.

These results are reminiscent of similar studies done in other PP5 enzymes in which mutational inactivation of the TPR domain resulted in an elevated basal activity and concomitant reduced response to unsaturated fatty acids [19]. Such studies led to the proposal that TPR domains regulate the PP5 catalytic region in a negative manner and that the interaction of arachidonic acid with TPR relieves this auto-inhibition. This was further supported by the demonstration that trypsin cleaved PP5 in the "hinge" sequence connecting the TPR region and catalytic domain, and that such cleavage produced a highly active enzyme which was refractory to further activation by fatty acids [18, 20, 21]. To test whether this is also true of PfPP5, we digested recombinant PfPP5 with trypsin, which resulted in the production of a trypsin-resistant fragment that was very similar in size to ΔTPR-PfPP5 (Fig. 3, lane 7). When assayed for phosphatase activity in vitro, the trypsin-digested PfPP5 indeed behaved like ΔTPR-PfPP5 in that it was about 3-fold more active than the full-length enzyme and was not activated by arachidonic acid any further (Fig. 5). Together these results suggest a role of PfPP5 TPR in auto-inhibition and in fatty acid-mediated activation.

Okadaic acid sensitivity of PfPP5

The differential sensitivity of the members of the PPP family to biological toxins has been exploited as a diagnostic tool in studying these phosphatases. The marine toxin okadaic acid (OA), in particular, inhibits the PP2A class of phosphatases at sub-nanomolar concentrations (IC50 ~ 0.05 nM) and the PP1 class in the low nanomolar range (IC50 ~ 5–45 nM). The SAPNYC motif of PPP enzymes is essential for OA-binding, and is found in all OA-sensitive phosphatases including PP1, PP2A, and PP5 [[25, 27, 28], and references therein]. This motif is also present in PfPP5 (Fig. 1). Phosphatases lacking this motif include the PP2B and PP2C classes, which are thus highly resistant to OA. PP2A contains an additional Cys residue two amino acids downstream (underlined in SAPNYCYRCG), believed to be important for the profound sensitivity of this class to OA [28], and this Cys is absent in PfPP5 and other PP5 enzymes (Fig. 1). As shown in Fig. 6, OA inhibited PfPP5 in a dose-dependent manner, with an IC₅₀ of 5.1 ± 0.2 nM. This moderate OA-sensitivity further establishes the identity of PfPP5, since other PP5 enzymes also exhibit an IC₅₀ value for OA in the 4–6 nM range [18]. The ΔTPR mutant exhibited an essentially identical IC₅₀ (not shown), supporting the notion that the TPR domains do not play a direct role in the interaction with OA.

PfPP5 binds Pf heat shock protein 90 (hsp90)

Hsp90 is a highly conserved molecular chaperone necessary for viability and for proper folding, processing, and function of proteins involved in several signal transduction pathways [31, 32]. Interestingly, mammalian hsp90 has been shown to specifically interact with PP5, and this interaction may have important consequences in PP5 structure and function [33, 34]. An hsp90 ortholog has been described in Pf [35], but its function is currently unknown. It was, therefore, logical for us to ask whether an interaction between PP5 and hsp90 also occurs in Pf. To initiate these studies, we performed a co-immunoprecipitation analysis of Pf extracts in which complexes containing PP5 were first precipitated with anti-PP5 antibody and the precipitate then probed with anti-hsp90 antibody in Western blot. Results (Fig. 7) show that Pfhsp90 is indeed found in complex with PfPP5, suggesting a potential role of this interaction in the mutual regulation of these two important proteins. The "control" non-immune serum did not react with either protein (lane 1) and did not precipitate hsp90 (lane 3).

PCR and other procedures

The following oligodeoxynucleotide primers were designed against the putative PfPP5 sequence found on chromosome 13 (AL049185; Plasmodium falciparum chr13_002073): ATGTTACACAACCATGATGTAGAAGAAG;TAAATGTTTTGATACAAATTATGAGC. These primers were used to amplify the PfPP5 gene or the cDNA using the Pf3D7 genomic DNA or Dd2 total RNA as template with PCR and RT-PCR, respectively. The RT reaction was performed using AMV reverse transcriptase, and all PCRs were performed with a 15:1 mixture of Taq:Pfu polymerase. The products were cloned into pGEM-T and confirmed by sequencing. The genomic and cDNA sequences were identical. This sequence was then subcloned into the XhoI and BamHI sites of the pET-15b vector by designing similar primers but containing the respective restriction sites. The TPR deletion mutant, starting at Glu-274 (bold in Fig. 1), was also amplified in a similar manner using specific primers, and then cloned. Introduction of the plasmids into E. coli BL21(DE3) cells containing the compatible RIG plasmid [39], induction of the proteins with IPTG, and purification through Ni-chelation chromatography were performed using the His-Bind kit reagents and procedure (Novagen).

Antibodies against PP5 and hsp90 were raised in rabbit and mouse, respectively, using standard procedures [40]. Fatty acids were purchased from Sigma (St. Louis, MO). SDS-PAGE and Western blot using chemiluminescence-based detection were performed essentially as described [15]. Immunoprecipitation of PfPP5 complexes followed by immunoblot analysis of the precipitate were carried out essentially as described [34].

Growth of Plasmodium and purification of PfPP5

P. falciparum (3D7 or Dd2) was grown in human A-positive erythrocytes and infected cells were harvested at about 15% parasitemia and lysed by the saponin method essentially as described [14, 15]. PfPP5 was partially purified using published procedures [18]. In brief, the parasites were resuspended in buffer A (50 mM Tris-Cl, pH 7.5, 10% glycerol, 1 mM DTT) plus 20 mM NaCl, 1% Triton X-100, leupeptin (1 μg/ml), and aprotinin (1 μg/ml) at an approximate protein concentration of 10 mg per ml. The suspension was sonicated and the lysate centrifuged at 100,000 × g for 30 min. The supernatant, called "total extract", was subjected to 45%–60% ammonium sulfate fractionation. The pellet was resuspended in 10 ml of buffer B (20 mM HEPES, pH 8.5, 1 mM EDTA, 1 mM DTT), desalted on a Sephadex G-25 column equilibrated with buffer B, then subjected to sequential chromatography on a 5 ml HiTrap heparin column (Pharmacia) and then a 5 ml HiTrap Q column (Pharmacia). PP5 activity was followed during the purification by assaying trypsin-stimulated pNPPase activity. The fractions were further confirmed by Western blot using anti-PfPP5 antibody. Parasite culture was synchronized essentially as described earlier [15].

Phosphatase assay

Phosphatase activity against pNPP or phosphohistone substrates was determined essentially as described previously. Standard assays were performed in buffer A containing 1 mM EDTA, and additional activators or inhibitors were added only where mentioned. Reactions were incubated at 37°C. ³²P-labeled phosphohistone was made as described [14, 15, 18]. When using phosphohistone as substrate, reactions were carried out in 80 μl followed by quantification of the liberated ³²Pi by phosphomolybdate extraction assay [14, 15, 18]. Reactions were followed with time, and results were corrected by subtraction of the corresponding values from an enzyme-free reaction. When p-nitrophenylphosphate was used as substrate (at 10 mM), the reaction was scaled up to 200 μl, and the liberated p-nitrophenol was quantitated by measurement of absorbance at 415 nm [14, 23]. Phosphatase reactions were routinely initiated with the addition of substrate. Dephosphorylation was kept to less than 10 % of the total phosphorylated substrate, and the reaction was linear with respect to enzyme concentration and time. When using potential activators (e.g., fatty acids) or inhibitors (e.g., okadaic acid), they were added to the enzyme mixture 10 minutes prior to initiating the reaction with the addition of substrate [14].